spec-preflight

Verifies tooling, authentication sessions, and credentials before autonomous implementation work.

Updated Mar 11, 2026
One-click install
npx skills add https://github.com/julien-m/livespec --skill spec-preflight-julien-m
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: spec-preflight
Source: https://github.com/julien-m/livespec/tree/main/.agent-sync/skills/spec-preflight
Command: npx skills add https://github.com/julien-m/livespec --skill spec-preflight-julien-m

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Before running autonomous implementation, teams need confidence that every required binary, OAuth session, and API token is present and valid. This Skill reads a versioned preflight manifest, executes all verification checks, auto-resolves what it can, and groups remaining human-required actions into a single blocking report. ## Core Features & Use Cases - Manifest-driven verification: Parses .specs/preflight.md with tooling, authentication, token, and custom check sections, then runs a 3-pass engine (parallel verify, sequential auto-resolve, grouped human blockers). - Stack-aware manifest generation: Regenerates the manifest from the project stack (_default.md), feature specs, and .env credential entries using a stack-to-checks catalog covering Cloudflare, Supabase, Vercel, Stripe, AWS, Playwright, and more. - Auto-fix and light modes: --fix installs missing tools via brew, cargo, npm, pip, or pipx and initializes simulators; --light re-verifies only expiring or new checks. - Use Case: Before launching an autonomous /spec-implement run on a Next.js + Supabase feature, run the preflight check to confirm node, supabase CLI, and stored tokens are valid, auto-install anything missing, and receive a READY verdict report. ## Quick Start Ask the AI to run the spec-preflight check to verify all tools, sessions, and tokens and write the readiness report before starting implementation.

Frequently Asked Questions about spec-preflight

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I verify required tools and credentials before running autonomous implementation?

Run the preflight check, which reads the `.specs/preflight.md` manifest and executes every verify command in parallel. It auto-resolves failures with install commands, groups remaining human actions, and writes a verdict report to `.specs/preflight-report.md`.

How do I generate a preflight manifest from my project stack?

Use the `--regenerate` flag to parse `.specs/stacks/_default.md`, match technologies against a built-in catalog (Cloudflare, Supabase, Vercel, Stripe, AWS, Playwright, and others), and scan feature specs plus `.env` for token checks. Custom sections between markers are preserved.

Can preflight automatically install missing tools?

Yes, the `--fix` flag dispatches installs through brew, cargo, npm, pip, or pipx and can initialize iOS simulators and Android AVDs. Checks marked `human` produce a numbered manual-action guide and exit code 2 instead of attempting installation.

What happens when a critical preflight check fails?

Any critical check still failing after the human-blocker pass blocks implementation entirely. Warning-severity failures only display a notice and allow the workflow to continue with a WARNINGS verdict.

Does the preflight report expose token or secret values?

No. Token checks only verify existence via `creds get` redirected to null, and the report records `exists` or `missing` for each entry. Secret values are never displayed in terminal output or the persisted report.